Colleagues,

Please find below the draft minutes of our meeting in Edinburgh. If anyone has 
any corrections, before 12:00 CEST on Friday 26th June, please let the 
Co-Chairs know, otherwise they will be posted to the WG website.

-----

Security Working Group Minutes RIPE 92
Tuesday, 19 May 2026 16:00 - 17:30 (UTC +1)
Chairs: Tobias Knecht, Brian Nisbet, Markus de Brün
Scribe: Ed Shryane
Status: Draft

The recordings and presentations are available at:
https://ripe92.ripe.net/programme/meeting-plan/sessions/79/

The stenography transcript is available at:
https://ripe92.ripe.net/programme/meeting-plan/sessions/79/transcript/

Security WG RIPE 92 Intro

Brian Nisbet welcomed everyone to the WG session. The minutes from RIPE 91 were 
approved as there were no comments.

LEA report 2025 and eEvidence implementation
Franca Bosompim, RIPE NCC

Presentation on the RIPE NCC’s LEA report 2025 and the implementation of 
eEvidence regulation.

Franca, Senior Legal Counsel for the RIPE NCC, presented the 2025 LEA 
transparency report. In 2025, the RIPE NCC received 99 requests from LEAs, 
eight of which were binding requests from Dutch LEAs. Most requests came from 
France and the U.S.

Dick Leaning, RIPE NCC, came on stage to expand further on the LEA Transparency 
report. The RIPE NCC’s Learning and Development department provided training to 
700 LEA staff and the goal is for LEAs to make requests which are in scope as 
this saves time for them and for the RIPE NCC.

Next Franca moved to the e-evidence package which is due to be transposed into 
Dutch law in 2027. This consists of the e-evidence regulation and directive. 
The regulation applies to the RIPE NCC, because it offers IP numbering services 
in the EU.

Subscriber data is e-evidence data that applies to the RIPE NCC. Traffic data 
or content data is not held by the RIPE NCC, because it doesn’t provide 
communication services. IP addresses are not explicitly included in the 
regulation, but as part of subscriber data. The RIPE NCC has objection rights 
to an order in certain cases. Requests must be kept confidential, and members 
are not notified of an order.

Alex de Joode, AMS-IX, said that Brussels had an opinion that the starting date 
would be 18 August, and if the Dutch were not ready, the RIPE NCC would also 
not be given extra time.

Leo Vegoda, PeeringDB, asked about the confidentiality requirement and whether 
reports would be added to future transparency reports.

Franca said that the RIPE NCC were investigating and would reply to the mailing 
list.

Alex de Joode further clarified that the amount of reports were not 
confidential, but the RIPE NCC would not yet know what exactly they would be 
allowed to report.

Residential Proxies and Infecting Infrastructure Working Session
Leslie Daigle

Leslie Daigle summarised a Global Cyber Alliance roundtable on the growing 
challenge of residential proxies and infected infrastructure. Attack traffic 
has increased significantly since October 2024, with residential proxies making 
malicious activity difficult to detect and mitigate. Participants agreed that 
traditional IP-blocking approaches did not scale, particularly with the 
widespread use of carrier-grade NAT. The workshop identified control-plane 
mitigation as a more promising approach, with next steps including a report 
outlining concrete actions and the development of a pilot project.

Merike Kaeo, Co-chair of the NetSec Working Group said they have been working 
on this topic for the past six to eight months, and that it would be useful to 
collaborate with FIRST, and ISPs globally who were part of that initiative and 
discussion.

Stephen Farrell, Trinity College Dublin, asked which major AI companies were 
the largest funders.

Leslie replied that she was not going to name them, but there were attempts to 
circumvent blocks on large-scale scraping of content.

Nils Trampel, University of Chemnitz, pointed out there was also an intentional 
use of residential proxies by users who sell their bandwidth, and they saw some 
abuse in a university setting from someone who unintentionally informed them of 
this use, and asked for more research into this.

Leslie agreed and asked what one does this with this information once 
identified.

Brian Nisbet replied that such services would break AUPs of Universities in the 
National Research Network.

The Moving Parts of Copyright and Cybercrime
Lee Kent, beIN

Lee Kent from beIN gave an update on what they have been doing over the past 12 
months and some of the research they had taken on board.

Non-compliant content providers cause a significant issue for rights holders. 
As part of that ecosystem they looked at the hosting infrastructure. By 
non-compliance they meant a no response to email or no action to comply with 
takedown requests from copyright owners. It is very much a small group of 
operators they have this narrowed down to.

He provided a breakdown of different types of hosting content, of which live TV 
streaming fairly down on the list at 3%, the majority of hosted content was 
phishing, malicious use and malware. beIN wanted legal clarity on the 
obligation of hosting providers to take down live content in real time.

On top of the digital services act, the EU was looking at a number of options 
to mitigate these risks. Critical issue of live piracy of audiovisual content. 
They are looking for collaboration and help across the Internet ecosystem.

There were no questions or comments.

The €30 attack box: inside the Android TV botnet ecosystem
Jérôme Meyer

Jérôme Meyer, a security researcher at Nokia, examined the rise of Android 
TV-based botnets and residential proxies, which now play a major role in 
large-scale DDoS attacks. He explained that low cost Android TV devices often 
arrive already compromised, allowing attackers to build botnets that can 
generate terabit-scale attacks and provide access to residential networks. 
Meyer stressed that operators should look beyond attack traffic and focus on 
detecting command-and-control activity. He concluded that network-edge 
disruption can be effective, but requires coordinated action from operators.

Lutz Donnerhacke, IKS Service GmbH, said that they had an ISP running numerous 
NAT devices and have had the same problems. They noticed that their devices 
were being broken and that they looked for outgoing packets without an active 
connection, and generated statistics which helped them track down devices in 
customer networks. He said he is interested to hear from others in how they 
detect such devices.

Jérôme replied that they have observed that a lot of proxy exit nodes were very 
chatty, and have a lot of long-lived TCP connections, and a combination of 
these two behaviours was a good way to detect proxy nodes.

Victor Guerrero, VGS, said they were suffering this on the server side. They 
had a large customer base and they had begun to see large traffic coming from 
Brazil since last August.

They blocked traffic from America but then the attacks started to come from 
Africa. He asked for ideas on how to detect this pattern before it hits the 
server. He said they could block traffic once there was at least one ping to 
the server and they were not attacking the infrastructure, but the server. He 
added that it can generate requests from 30,000 IPs in less than 5 minutes. 
It’s impossible to deal with that in an effective way, and that he is open to 
ideas.

Jérôme replied that they should not single out Brazil and although it’s the top 
source of traffic, it was really from everywhere. It was trivial to rotate 
between tens of thousands of IP addresses and there had been efforts to 
characterise the traffic that was proxied through those nodes. He added that 
there may be a way to have a signal do some detection.

Trust but Verify: An Assessment of Vulnerability Tagging Services
Szu-Chun Huang, TU Delft

Szu-Chun Huang presented research evaluating the accuracy of three Internet 
vulnerability tagging services. The study compared reports against independent 
vulnerability scans and controlled testing using intentionally vulnerable 
systems. The results revealed significant discrepancies between the services, 
raising concerns about the completeness and reliability of vulnerability 
reports used to assess Internet security risks.

There were no questions or comments.

Security WG RIPE 92 Outro
Brian Nisbet

A.O.B.

There was no other business.

Agenda for RIPE 93

Brian then closed the session and encouraged the audience to use the Security 
Working Group mailing list for discussion.



Brian Nisbet (he/him)
Head of Service Operations
Asiera (formerly HEAnet), Ireland's National Research and Education Network
North Dock Two, 93-94 North Wall Quay, Dublin 1, D01 V8Y6
+35316609040 [email protected] www.heanet.ie<http://www.heanet.ie> 
www.asiera.ie<https://www.asiera.ie/>
Registered in Ireland, No. 275301. CRA No. 20036270
-----
To unsubscribe from this mailing list or change your subscription options, 
please visit: https://mailman.ripe.net/mailman3/lists/security-wg.ripe.net/
As we have migrated to Mailman 3, you will need to create an account with the 
email matching your subscription before you can change your settings. 
More details at: https://www.ripe.net/membership/mail/mailman-3-migration/

Reply via email to